# Human Record — Hannibal through surviving evidence Status: **PUBLIC RECORD 4 / HISTORICAL-PERSON SOURCE-SURVIVAL / PUBLIC SOURCES / NOT BIOGRAPHICAL CANON / OPEN TO CORRECTION** Record ID: `hannibal-source-survival` Record version: **0.1.3** Recorded: 18 September 2026. ## Purpose This record asks a narrow Human Record question: > What is the smallest inspectable evidence path by which the historical human conventionally called Hannibal Barca remains reachable now, without turning surviving literary tradition into the person himself? The record does **not** attempt a definitive biography, settle the Alpine route, reconstruct Hannibal's appearance, rank commanders, or treat every later retelling as an independent witness. ```text NAME != ENTITY ENTITY != BIOGRAPHY ENTITY != CLAIMS ABOUT ENTITY SURVIVING_SOURCE != COMPLETE_HISTORY ATTESTED_WORK != INSPECTED_SOURCE REPRESENTATION_OF_ENTITY != EVIDENCE_OF_ENTITY_APPEARANCE SOURCE_REPORTS_X != THR_DIRECTLY_OBSERVED_X ``` ## Thin entity anchor Working entity: **Hannibal Barca**, the historical human named as Hannibal in the checked ancient literary surfaces below. The entity anchor deliberately carries no exact birth/death date, portrait, Alpine pass, childhood-oath wording, or full office/kinship profile. Those are assertions requiring their own evidence paths. The opaque candidate entity ID is a routing handle only. It is not evidence. The full display string **Hannibal Barca** is used here as a conventional modern routing label. The checked Nepos Latin passage attests the literal name **Hannibal**. The checked Polybius English surface renders the name as **Hannibal**. A separate Perseus passage-vocabulary index for **3.33.17–18** associates both sections with the lemma **Ἀννίβας**. That is lemma-level indexing, not direct inspection of the inflected Greek source-text spelling. It therefore does not by itself narrow the literal spelling to a particular form, establish identity, or inspect a manuscript witness. ```text CONVENTIONAL_DISPLAY_LABEL != SOURCE_ATTESTED_STRING ``` ## Why this case was selected Hannibal was selected deliberately as a **historical-person / source-survival stress case** after the first three public Human Record entries had already tested artwork provenance, repeated-claim ancestry, and living-practice transmission. The selection-time candidate pool and alternatives considered were not preserved from the pre-publication selection process. That gap does not establish that no alternatives were considered; it means this record cannot honestly reconstruct them after the fact. Selection does not establish representativeness, historical priority, importance, or entitlement to become record 4. ```text MISSING_SELECTION_HISTORY != PERMISSION_TO_INVENT_IT SELECTED_AS_STRESS_CASE != REPRESENTATIVE_SAMPLE != MOST_IMPORTANT_PERSON != RECORD_4_ENTITLEMENT ``` ## Checked source apertures ### Polybius, *Histories* Book 3 Public reading surface: `https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Polybius/3*.html` The host presents the English text as the 1922 Loeb translation and warns that the web page has not yet been proofread. THR therefore treats it as a convenient modern reading surface for bounded passage inspection, **not** as a manuscript witness or a checked critical edition. This record uses four bounded passages: - **3.20.5** — Polybius names Chaereas and Sosylus and dismisses their authority sharply. This establishes that Polybius knew a Sosylus tradition; it does not make Polybius' evaluation neutral or final. - **3.33.17–18** — Polybius says his unusually detailed troop figures came from a bronze tablet on the Lacinian promontory on which Hannibal had made out the lists. THR has not observed that tablet and does not convert Polybius' report into a surviving inscription object. - **3.48.12** — Polybius says he inquired from men present at the crossing and personally inspected the country and crossed the Alps himself. This is evidence about Polybius' claimed method, not automatic proof of every detail in his reconstruction. - **3.56.3–4** — Polybius reports Hannibal's post-crossing force totals and attributes them to an inscription/column at Lacinium. This record does **not** establish that the carrier described here is physically identical to the bronze tablet described at 3.33.17–18. #### Greek vocabulary-index check at 3.33.17–18 Perseus identifies a Greek edition of Polybius as `urn:cts:greekLit:tlg0543.tlg001.perseus-grc2`, based on the Büttner-Wobst edition. THR inspected the passage-specific vocabulary/lemma lists for **3.33.17** and **3.33.18**. Each list associates its section with the lemma **Ἀννίβας** once. Routes: - https://vocab.perseus.org/word-list/urn:cts:greekLit:tlg0543.tlg001.perseus-grc2:3.33.17/ - https://vocab.perseus.org/word-list/urn:cts:greekLit:tlg0543.tlg001.perseus-grc2:3.33.18/ This is a bounded vocabulary-index check of the named sections, not direct inspection of the inflected Greek source-text spelling, a manuscript inspection, or a complete Greek-text audit. The Scaife full-reader route was not available through this aperture, and the other THR-cited Polybius passages were not upgraded by this check. ```text VOCABULARY LEMMA != SOURCE-LITERAL SPELLING LEMMA ASSOCIATION != HISTORICAL IDENTITY PROVED PASSAGE INDEX != MANUSCRIPT WITNESS ONE INDEX CHECK != WHOLE BOOK VERIFIED ``` ### Cornelius Nepos, *Life of Hannibal* 13.3 Public Latin/commentary surface: `https://dcc.dickinson.edu/nepos-hannibal/chapter-13` Nepos reports that Silenus and Sosylus lived with Hannibal in camp and wrote about the war, and identifies Sosylus as Hannibal's teacher of Greek literature. This record records that **Nepos reports this relationship**. It does not create Silenus' or Sosylus' lost/fragmentary works as inspected THR sources merely because later authors attest them. ## Source-survival pressure These two checked texts do not establish two independent witnesses to the same event. They support different attributed propositions; their number is not a corroboration count. This record does not establish direct dependence between Nepos and Polybius either. Polybius is both: - a surviving narrative source for Hannibal's campaign; and - a source critic who names and rejects Sosylus. Nepos, writing later, attests Hannibal-entourage historians whose works are not preserved here as complete inspected source objects. Modern scholarship also treats later Hannibal narratives as source-dependent. A current source-criticism orientation checked for this record is the Cambridge *Libyan Studies* article “Generals and judges: command, constitution and the fate of Carthage”: `https://www.cambridge.org/core/journals/libyan-studies/article/generals-and-judges-command-constitution-and-the-fate-of-carthage/757F46BAE0CA1A08373A15D1E497198F` It notes Polybius' use of Carthaginian documentary/informant material, Sosylus in Hannibal's entourage, and heavy later reliance on Polybius while preserving the existence of additional sources and interpretive lenses. This is orientation for source ancestry, not a substitute for reading the ancient passages. Livy is therefore **not counted as an independent corroborating witness by default** in this record. A later source can be added when its relation to earlier material is itself evidenced. ```text LATER_SOURCE != INDEPENDENT_WITNESS SOURCE_DEPENDENCE_UNKNOWN != SOURCE_INDEPENDENT ONE_SIDED_BY_SURVIVAL != FALSE MISSING_CARTHAGINIAN_APERTURE != CARTHAGINIAN_HISTORY_DID_NOT_EXIST ``` The current checked narrative aperture is Greek/Roman-mediated. That asymmetry is part of the record, not a footnote to hide. ## Source recoverability — Nepos edition A preservation check on 20 September 2026 did **not** establish an archived copy of the exact Dickinson College Commentaries chapter-13 web representation used by this record. It did establish stronger institutional recovery routes for the closely related published Dickinson College Commentaries edition by Bret Mulligan: - Open Book Publishers, *Cornelius Nepos, Life of Hannibal: Latin Text, Notes, Maps, Illustrations and Vocabulary*, DOI `10.11647/OBP.0068`, identifies the book as Dickinson College Commentaries volume 1 and exposes the Nepos text and notes; - the Library of Congress provides an unrestricted digital copy under LCCN `2019467885` / persistent handle `hdl.loc.gov/loc.gdc/gdcebookspublic.2019467885`. These routes improve recoverability of the edition-level text/notes if the live DCC page changes or disappears. They were not compared byte-for-byte against the observed chapter page and do not preserve its exact layout, hyperlinks, commentary state or web response. ```text RELATED EDITION RECOVERABLE != EXACT WEB REPRESENTATION PRESERVED PRESERVATION ROUTE != HISTORICAL TRUTH INSTITUTIONAL COPY != BYTE-IDENTICAL DCC PAGE ``` No historical assertion or source interpretation changed in this maintenance revision. ## Source recoverability — Polybius edition A preservation check on 25 September 2026 followed the printed-edition route named by the University of Chicago / LacusCurtius Book III reading surface. Internet Archive item `historieswitheng02poly`, contributed from the New York Public Library, currently identifies: - Polybius, *The Histories*, **Volume 2**; - W. R. Paton; - Cambridge: Harvard University Press; London: Heinemann; - publication date **[1922]**; - Greek and English on opposite pages. This establishes a public institutional recovery route for a Volume II carrier in the printed-edition family used by the reading surface. The item metadata does **not** by itself establish whether the scanned physical copy is the original printing or a later reprint, and THR has not compared the scan against the Thayer HTML for textual or byte identity. ```text PRINTED EDITION ROUTE IDENTIFIED != EXACT WEB REPRESENTATION PRESERVED ITEM METADATA != PHYSICAL PRINTING STATE ARCHIVE SCAN != THAYER TRANSCRIPTION PRESERVATION ROUTE != HISTORICAL TRUTH ``` No historical assertion or source interpretation changed in this maintenance revision. ## Public machine route The public machine record is: `https://thehumanrecord.net/cases/hannibal-barca.json` It keeps two source-literal `Hannibal` mentions separate — one from Nepos and one from Polybius — and candidate-links both to the same thin entity anchor. That allows later evidence to split or revise the mappings without rewriting either source literal. ```text SAME_LITERAL_ACROSS_SOURCES != IDENTITY_PROVED CANDIDATE_LINK != ENTITY_CERTAINTY ``` The earlier research fixture remains in the repository as build/test lineage. It is not the public machine record. ## What the executable fixture represents The preserved pre-publication research fixture contains only: - one candidate person entity; - one literal Hannibal mention; - two inspected web reading surfaces; - source-attributed assertions; - no direct-observation assertion; - no invented Sosylus/Silenus source object; - no catalogue identity. It deliberately tests the current production validators after the 18 September semantic repairs. Default/public validation continues to reject the research fixture's fabricated unpublished `record_id`; the public machine record uses `hannibal-source-survival`. Explicit isolated research validation may omit `record_id` while retaining source + owned observation + context. That is workflow permission, not privacy or publication. ## Current bounded findings 1. The existing entity/source/assertion envelopes can carry this thin historical-human record without adding a new entity type or truth scale. 2. A source can report a claim about Hannibal without the entity anchor absorbing that claim as identity. 3. A later source can attest an earlier source tradition without THR pretending to have inspected the earlier work. 4. Polybius' claimed investigative method and documentary source can be recorded as **claims by Polybius** without becoming direct THR observations. 5. The current source set is asymmetrical and incomplete. The record shows that asymmetry rather than multiplying later repetitions. None of these findings establishes the historical truth of every underlying event. ## What remains unknown or deliberately unclaimed This record does not establish: - exact birth or death dates; - the exact Alpine pass; - a complete source genealogy between Polybius, Livy and other surviving writers; - the contents of all surviving Sosylus/Silenus fragments; - the full Punic/Carthaginian documentary record; - Hannibal's appearance; - the exact wording/historicity of the childhood oath; - complete archaeological, numismatic or epigraphic evidence; - a current scholarly consensus on every disputed episode. ```text NOT_CHECKED != ABSENT UNKNOWN != EMPTY_FIELD_TO_FILL NO_SURVIVING_WHOLE != NO_FRAGMENT ``` ## Publication boundary Publication makes this a fourth Human Record encounter. It does not establish: - reader benefit; - historical authority over specialist scholarship; - completeness of the source genealogy; - identity certainty merely because the record has an entity handle. The record was exposed as a non-catalogued reader proposal before promotion. No reader-specific outside falsifier arrived in that interval. That null result is recorded as a null result, not as approval. ```text PUBLICATION != VALIDATION NO_OUTSIDE_READER_REPLY != ENDORSEMENT RECORD_4 != FOURTH_MOST_IMPORTANT_HUMAN_RECORD ``` ## Strong-owner boundary THR does not need a new historical ontology for this case. Relevant owners already include: - CIDOC CRM family for cultural-heritage/event semantics; - W3C PROV for provenance/derivation; - CRMinf for argumentation/inference provenance; - Wikibase-style qualified statements/references/revision history; - classical citation and historical source criticism for text identity and dependence. The Human Record residue being tested is smaller: make source ancestry, unknowns, disagreement and correction visible to ordinary human and machine readers without pretending the sparse graph is the world. ## Publication decision Before promotion, the candidate was required to preserve direct-source locators, show the Greek/Roman survival asymmetry, survive source-dependence and biography-fact attacks, retain a correction route and explicit unexamined material, and remain useful without claiming to outperform specialist historical owners. Those checks justified publication as a fourth THR encounter. They did not establish historical truth, reader benefit, or a general method advantage. ```text STRUCTURAL_PASS != HISTORICAL_TRUTH PUBLICATION != VALIDATION PUBLICATION_DECISION != OWNER_SUPERIORITY THR_SURVIVAL != PURPOSE ``` ## Public routes - Human view: https://thehumanrecord.net/records/hannibal.html - Machine record: https://thehumanrecord.net/cases/hannibal-barca.json - Catalogue: https://thehumanrecord.net/records/catalog.json - Correction route: https://thehumanrecord.net/CONTRIBUTE.md The repository history preserves the earlier candidate state and the issue #40 identity-model repairs that preceded publication. ```text PUBLIC_RECORD != HISTORICAL_ORACLE CORRECTED_RECORD != ORIGINAL_EVENT_CHANGED ```